Laine Stump [Thu, 1 May 2014 08:40:41 +0000 (11:40 +0300)]
qemu: fix crash when removing <filterref> from interface with update-device
If a domain network interface that contains a <filterref> is modified
"live" using "virsh update-device --live", libvirtd would crash. This
was because the code supporting live update of an interface's
filterref was assuming that a filterref might be added or modified,
but didn't account for removing the filterref, resulting in a null
dereference of the filter name.
Introduced with commit 258fb278, which was first in libvirt v1.0.1.
This addresses https://bugzilla.redhat.com/show_bug.cgi?id=1093301
qemu: make sure agent returns error when required data are missing
Commit 5b3492fa aimed to fix this and caught one error but exposed
another one. When agent command is being executed and the thread
waiting for the reply is woken up by an event (e.g. EOF in case of
shutdown), the command finishes with no data (rxObject == NULL), but
no error is reported, since this might be desired by the caller
(e.g. suspend through agent). However, in other situations, when the
data are required (e.g. getting vCPUs), we proceed to getting desired
data out of the reply, but none of the virJSON*() functions works well
with NULLs. I chose the way of a new parameter for qemuAgentCommand()
function that specifies whether reply is required and behaves
according to that.
On all the places where qemuAgentComand() was called, we did a check
for errors in the reply. Unfortunately, some of the places called
qemuAgentCheckError() without checking for non-null reply which might
have resulted in a crash.
So this patch makes the error-checking part of qemuAgentCommand()
itself, which:
a) makes it look better,
b) makes the check mandatory and, most importantly,
c) checks for the errors if and only if it is appropriate.
This actually fixes a potential crashers when qemuAgentComand()
returned 0, but reply was NULL. Having said that, it *should* fix the
following bug:
Michal Privoznik [Wed, 19 Mar 2014 17:10:34 +0000 (18:10 +0100)]
virNetClientSetTLSSession: Restore original signal mask
Currently, we use pthread_sigmask(SIG_BLOCK, ...) prior to calling
poll(). This is okay, as we don't want poll() to be interrupted.
However, then - immediately as we fall out from the poll() - we try to
restore the original sigmask - again using SIG_BLOCK. But as the man
page says, SIG_BLOCK adds signals to the signal mask:
SIG_BLOCK
The set of blocked signals is the union of the current set and the set argument.
Therefore, when restoring the original mask, we need to completely
overwrite the one we set earlier and hence we should be using:
SIG_SETMASK
The set of blocked signals is set to the argument set.
If a VM dies very early during an attempted connect to the guest agent
while the locks are down the domain monitor object will be freed. The
object is then accessed later as any failure during guest agent startup
isn't considered fatal.
In the current upstream version this doesn't lead to a crash as
virObjectLock called when entering the monitor in
qemuProcessDetectVcpuPIDs checks the pointer before attempting to
dereference (lock) it. The NULL pointer is then caught in the monitor
helper code.
Before the introduction of virObjectLockable - observed on 0.10.2 - the
pointer is locked directly via virMutexLock leading to a crash.
To avoid this problem we need to differentiate between the guest agent
not being present and the VM quitting when the locks were down. The fix
reorganizes the code in qemuConnectAgent to add the check and then adds
special handling to the callers.
The nwfilter conf update mutex previously serialized
updates to the internal data structures for firewall
rules, and updates to the firewall itself. The latter
was recently turned into a read/write lock, and filter
instantiation allowed to proceed in parallel. It was
believed that this was ok, since each filter is created
on a separate iptables/ebtables chain.
It turns out that there is a subtle lock ordering problem
on virNWFilterObjPtr instances. __virNWFilterInstantiateFilter
will hold a lock on the virNWFilterObjPtr it is instantiating.
This in turn invokes virNWFilterInstantiate which then invokes
virNWFilterDetermineMissingVarsRec which then invokes
virNWFilterObjFindByName. This iterates over every single
virNWFilterObjPtr in the list, locking them and checking their
name. So if 2 or more threads try to instantiate a filter in
parallel, they'll all hold 1 lock at the top level in the
__virNWFilterInstantiateFilter method which will cause the
other thread to deadlock in virNWFilterObjFindByName.
The fix is to add an exclusive mutex to serialize the
execution of __virNWFilterInstantiateFilter.
CVE-2013-6456: Avoid unsafe use of /proc/$PID/root in LXC hotunplug code
Rewrite multiple hotunplug functions to to use the
virProcessRunInMountNamespace helper. This avoids
risk of a malicious guest replacing /dev with an absolute
symlink, tricking the driver into changing the host OS
filesystem.
CVE-2013-6456: Avoid unsafe use of /proc/$PID/root in LXC chardev hostdev hotplug
Rewrite lxcDomainAttachDeviceHostdevMiscLive function
to use the virProcessRunInMountNamespace helper. This avoids
risk of a malicious guest replacing /dev with a absolute
symlink, tricking the driver into changing the host OS
filesystem.
CVE-2013-6456: Avoid unsafe use of /proc/$PID/root in LXC block hostdev hotplug
Rewrite lxcDomainAttachDeviceHostdevStorageLive function
to use the virProcessRunInMountNamespace helper. This avoids
risk of a malicious guest replacing /dev with a absolute
symlink, tricking the driver into changing the host OS
filesystem.
CVE-2013-6456: Avoid unsafe use of /proc/$PID/root in LXC USB hotplug
Rewrite lxcDomainAttachDeviceHostdevSubsysUSBLive function
to use the virProcessRunInMountNamespace helper. This avoids
risk of a malicious guest replacing /dev with a absolute
symlink, tricking the driver into changing the host OS
filesystem.
CVE-2013-6456: Avoid unsafe use of /proc/$PID/root in LXC disk hotplug
Rewrite lxcDomainAttachDeviceDiskLive function to use the
virProcessRunInMountNamespace helper. This avoids risk of
a malicious guest replacing /dev with a absolute symlink,
tricking the driver into changing the host OS filesystem.
Eric Blake [Tue, 24 Dec 2013 05:55:51 +0000 (22:55 -0700)]
CVE-2013-6456: Avoid unsafe use of /proc/$PID/root in LXC shutdown/reboot code
Use helper virProcessRunInMountNamespace in lxcDomainShutdownFlags and
lxcDomainReboot. Otherwise, a malicious guest could use symlinks
to force the host to manipulate the wrong file in the host's namespace.
Idea by Dan Berrange, based on an initial report by Reco
<recoverym4n@gmail.com> at
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=732394
Add helper for running code in separate namespaces
Implement virProcessRunInMountNamespace, which runs callback of type
virProcessNamespaceCallback in a container namespace. This uses a
child process to run the callback, since you can't change the mount
namespace of a thread. This implies that callbacks have to be careful
about what code they run due to async safety rules.
Idea by Dan Berrange, based on an initial report by Reco
<recoverym4n@gmail.com> at
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=732394
Signed-off-by: Daniel Berrange <berrange@redhat.com> Signed-off-by: Eric Blake <eblake@redhat.com>
(cherry picked from commit 7c72ef6f555f1f9844d51be2f38f078bc908652c)
Add a helper function which takes a file path and ensures
that all directory components leading up to the file exist.
IOW, it strips the filename part of the path and passes
the result to virFileMakePath.
Move check for cgroup devices ACL upfront in LXC hotplug
The check for whether the cgroup devices ACL is available is
done quite late during LXC hotplug - in fact after the device
node is already created in the container in some cases. Better
to do it upfront so we fail immediately.
Fix reset of cgroup when detaching USB device from LXC guests
When detaching a USB device from an LXC guest we must remove
the device from the cgroup ACL. Unfortunately we were telling
the cgroup code to use the guest /dev path, not the host /dev
path, and the guest device node had already been unlinked.
This was, however, fortunate since the code passed &priv->cgroup
instead of priv->cgroup, so would have crash if the device node
were accessible.
The LXC code missed the 'usb' component out of the path
/dev/bus/usb/$BUSNUM/$DEVNUM, so it failed to actually
setup cgroups for the device. This was in fact lucky
because the call to virLXCSetupHostUsbDeviceCgroup
was also mistakenly passing '&priv->cgroup' instead of
just 'priv->cgroup'. So once the path is fixed, libvirtd
would then crash trying to access the bogus virCgroupPtr
pointer. This would have been a security issue, were it
not for the bogus path preventing the pointer reference
being reached.
virDomainDefCompatibleDevice blocks use of USB if no USB
controller is present. This is not correct for containers
since devices can be assigned directly regardless of any
controllers.
Commit f9f56340 for CVE-2014-0028 almost had the right idea - we
need to check the ACL rules to filter which events to send. But
it overlooked one thing: the event dispatch queue is running in
the main loop thread, and therefore does not normally have a
current virIdentityPtr. But filter checks can be based on current
identity, so when libvirtd.conf contains access_drivers=["polkit"],
we ended up rejecting access for EVERY event due to failure to
look up the current identity, even if it should have been allowed.
Furthermore, even for events that are triggered by API calls, it
is important to remember that the point of events is that they can
be copied across multiple connections, which may have separate
identities and permissions. So even if events were dispatched
from a context where we have an identity, we must change to the
correct identity of the connection that will be receiving the
event, rather than basing a decision on the context that triggered
the event, when deciding whether to filter an event to a
particular connection.
If there were an easy way to get from virConnectPtr to the
appropriate virIdentityPtr, then object_event.c could adjust the
identity prior to checking whether to dispatch an event. But
setting up that back-reference is a bit invasive. Instead, it
is easier to delay the filtering check until lower down the
stack, at the point where we have direct access to the RPC
client object that owns an identity. As such, this patch ends
up reverting a large portion of the framework of commit f9f56340.
We also have to teach 'make check' to special-case the fact that
the event registration filtering is done at the point of dispatch,
rather than the point of registration. Note that even though we
don't actually use virConnectDomainEventRegisterCheckACL (because
the RegisterAny variant is sufficient), we still generate the
function for the purposes of documenting that the filtering
takes place.
Also note that I did not entirely delete the notion of a filter
from object_event.c; I still plan on using that for my upcoming
patch series for qemu monitor events in libvirt-qemu.so. In
other words, while this patch changes ACL filtering to live in
remote.c and therefore we have no current client of the filtering
in object_event.c, the notion of filtering in object_event.c is
still useful down the road.
* src/check-aclrules.pl: Exempt event registration from having to
pass checkACL filter down call stack.
* daemon/remote.c (remoteRelayDomainEventCheckACL)
(remoteRelayNetworkEventCheckACL): New functions.
(remoteRelay*Event*): Use new functions.
* src/conf/domain_event.h (virDomainEventStateRegister)
(virDomainEventStateRegisterID): Drop unused parameter.
* src/conf/network_event.h (virNetworkEventStateRegisterID):
Likewise.
* src/conf/domain_event.c (virDomainEventFilter): Delete unused
function.
* src/conf/network_event.c (virNetworkEventFilter): Likewise.
* src/libxl/libxl_driver.c: Adjust caller.
* src/lxc/lxc_driver.c: Likewise.
* src/network/bridge_driver.c: Likewise.
* src/qemu/qemu_driver.c: Likewise.
* src/remote/remote_driver.c: Likewise.
* src/test/test_driver.c: Likewise.
* src/uml/uml_driver.c: Likewise.
* src/vbox/vbox_tmpl.c: Likewise.
* src/xen/xen_driver.c: Likewise.
The NWFilter code has as a deadlock race condition between
the virNWFilter{Define,Undefine} APIs and starting of guest
VMs due to mis-matched lock ordering.
In the virNWFilter{Define,Undefine} codepaths the lock ordering
is
This has the effect of serializing VM startup once again, even if
no nwfilters are applied to the guest. There is also the possibility
of deadlock due to a call graph loop via virNWFilterInstantiate
and virNWFilterInstantiateFilterLate.
These two problems mean the lock must be turned into a read/write
lock instead of a plain mutex at the same time. The lock is used to
serialize changes to the "driver->nwfilters" hash, so the write lock
only needs to be held by the define/undefine methods. All other
methods can rely on a read lock which allows good concurrency.
Eric Blake [Tue, 14 Jan 2014 17:29:34 +0000 (10:29 -0700)]
event: filter global events by domain:getattr ACL [CVE-2014-0028]
Ever since ACL filtering was added in commit 7639736 (v1.1.1), a
user could still use event registration to obtain access to a
domain that they could not normally access via virDomainLookup*
or virConnectListAllDomains and friends. We already have the
framework in the RPC generator for creating the filter, and
previous cleanup patches got us to the point that we can now
wire the filter through the entire object event stack.
Furthermore, whether or not domain:getattr is honored, use of
global events is a form of obtaining a list of networks, which
is covered by connect:search_domains added in a93cd08 (v1.1.0).
Ideally, we'd have a way to enforce connect:search_domains when
doing global registrations while omitting that check on a
per-domain registration. But this patch just unconditionally
requires connect:search_domains, even when no list could be
obtained, based on the following observations:
1. Administrators are unlikely to grant domain:getattr for one
or all domains while still denying connect:search_domains - a
user that is able to manage domains will want to be able to
manage them efficiently, but efficient management includes being
able to list the domains they can access. The idea of denying
connect:search_domains while still granting access to individual
domains is therefore not adding any real security, but just
serves as a layer of obscurity to annoy the end user.
2. In the current implementation, domain events are filtered
on the client; the server has no idea if a domain filter was
requested, and must therefore assume that all domain event
requests are global. Even if we fix the RPC protocol to
allow for server-side filtering for newer client/server combos,
making the connect:serach_domains ACL check conditional on
whether the domain argument was NULL won't benefit older clients.
Therefore, we choose to document that connect:search_domains
is a pre-requisite to any domain event management.
Network events need the same treatment, with the obvious
change of using connect:search_networks and network:getattr.
Eric Blake [Tue, 14 Jan 2014 00:30:59 +0000 (17:30 -0700)]
Fix memory leak in virObjectEventCallbackListRemoveID()
While running objecteventtest, it was found that valgrind pointed out the
following memory leak:
==13464== 5 bytes in 1 blocks are definitely lost in loss record 7 of 134
==13464== at 0x4A0887C: malloc (vg_replace_malloc.c:270)
==13464== by 0x341F485E21: strdup (strdup.c:42)
==13464== by 0x4CAE28F: virStrdup (virstring.c:554)
==13464== by 0x4CF3CBE: virObjectEventCallbackListAddID (object_event.c:286)
==13464== by 0x4CF49CA: virObjectEventStateRegisterID (object_event.c:729)
==13464== by 0x4CF73FE: virDomainEventStateRegisterID (domain_event.c:1424)
==13464== by 0x4D7358F: testConnectDomainEventRegisterAny (test_driver.c:6032)
==13464== by 0x4D600C8: virConnectDomainEventRegisterAny (libvirt.c:19128)
==13464== by 0x402409: testDomainStartStopEvent (objecteventtest.c:232)
==13464== by 0x403451: virtTestRun (testutils.c:138)
==13464== by 0x402012: mymain (objecteventtest.c:395)
==13464== by 0x403AF2: virtTestMain (testutils.c:593)
==13464==
When writing commit 173c291, I missed the fact virNetServerClientClose
unlocks the client object before actually clearing client->sock and thus
it is possible to hit a window when client->keepalive is NULL while
client->sock is not NULL. I was thinking client->sock == NULL was a
better check for a closed connection but apparently we have to go with
client->keepalive == NULL to actually fix the crash.
When a client closes its connection to libvirtd early during
virConnectOpen, more specifically just after making
REMOTE_PROC_CONNECT_SUPPORTS_FEATURE call to check if
VIR_DRV_FEATURE_PROGRAM_KEEPALIVE is supported without even waiting for
the result, libvirtd may crash due to a race in keep-alive
initialization. Once receiving the REMOTE_PROC_CONNECT_SUPPORTS_FEATURE
call, the daemon's event loop delegates it to a worker thread. In case
the event loop detects EOF on the connection and calls
virNetServerClientClose before the worker thread starts to handle
REMOTE_PROC_CONNECT_SUPPORTS_FEATURE call, client->keepalive will be
disposed by the time virNetServerClientStartKeepAlive gets called from
remoteDispatchConnectSupportsFeature. Because the flow is common for
both authenticated and read-only connections, even unprivileged clients
may cause the daemon to crash.
To avoid the crash, virNetServerClientStartKeepAlive needs to check if
the connection is still open before starting keep-alive protocol.
Every libvirt release since 0.9.8 is affected by this bug.
Jiri Denemark [Fri, 20 Dec 2013 13:50:02 +0000 (14:50 +0100)]
qemu: Avoid using stale data in virDomainGetBlockInfo
CVE-2013-6458
Generally, every API that is going to begin a job should do that before
fetching data from vm->def. However, qemuDomainGetBlockInfo does not
know whether it will have to start a job or not before checking vm->def.
To avoid using disk alias that might have been freed while we were
waiting for a job, we use its copy. In case the disk was removed in the
meantime, we will fail with "cannot find statistics for device '...'"
error message.
When virDomainDetachDeviceFlags is called concurrently to
virDomainBlockStats: libvirtd may crash because qemuDomainBlockStats
finds a disk in vm->def before getting a job on a domain and uses the
disk pointer after getting the job. However, the domain in unlocked
while waiting on a job condition and thus data behind the disk pointer
may disappear. This happens when thread 1 runs
virDomainDetachDeviceFlags and enters monitor to actually remove the
disk. Then another thread starts running virDomainBlockStats, finds the
disk in vm->def, and while it's waiting on the job condition (owned by
the first thread), the first thread finishes the disk removal. When the
second thread gets the job, the memory pointed to be the disk pointer is
already gone.
That said, every API that is going to begin a job should do that before
fetching data from vm->def.
Dario Faggioli [Fri, 20 Dec 2013 15:29:47 +0000 (16:29 +0100)]
libxl: avoid crashing if calling `virsh numatune' on inactive domain
by, in libxlDomainGetNumaParameters(), calling libxl_bitmap_init() as soon as
possible, which avoids getting to 'cleanup:', where libxl_bitmap_dispose()
happens, without having initialized the nodemap, and hence crashing after some
invalid free()-s:
Signed-off-by: Dario Faggili <dario.faggioli@citrix.com> Cc: Jim Fehlig <jfehlig@suse.com> Cc: Ian Jackson <Ian.Jackson@eu.citrix.com>
(cherry picked from commit f9ee91d35510ccbc6fc42cef8864b291b2d220f4)
The function doesn't check whether the request is made for active or
inactive domain. Thus when the domain is not running it still tries
accessing non-existing cgroups (priv->cgroup, which is NULL).
I re-made the function in order for it to work the same way it's qemu
counterpart does.
Reproducer:
1) Define an LXC domain
2) Do 'virsh memtune <domain> --hard-limit 133T'
Backtrace:
Thread 6 (Thread 0x7fffec8c0700 (LWP 26826)):
#0 0x00007ffff70edcc4 in virCgroupPathOfController (group=0x0, controller=3,
key=0x7ffff75734bd "memory.limit_in_bytes", path=0x7fffec8bf718) at util/vircgroup.c:1764
#1 0x00007ffff70e9206 in virCgroupSetValueStr (group=0x0, controller=3,
key=0x7ffff75734bd "memory.limit_in_bytes", value=0x7fffe409f360 "1073741824")
at util/vircgroup.c:669
#2 0x00007ffff70e98b4 in virCgroupSetValueU64 (group=0x0, controller=3,
key=0x7ffff75734bd "memory.limit_in_bytes", value=1073741824) at util/vircgroup.c:740
#3 0x00007ffff70ee518 in virCgroupSetMemory (group=0x0, kb=1048576) at util/vircgroup.c:1904
#4 0x00007ffff70ee675 in virCgroupSetMemoryHardLimit (group=0x0, kb=1048576)
at util/vircgroup.c:1944
#5 0x00005555557d54c8 in lxcDomainSetMemoryParameters (dom=0x7fffe40cc420,
params=0x7fffe409f100, nparams=1, flags=0) at lxc/lxc_driver.c:774
#6 0x00007ffff72c20f9 in virDomainSetMemoryParameters (domain=0x7fffe40cc420,
params=0x7fffe409f100, nparams=1, flags=0) at libvirt.c:4051
#7 0x000055555561365f in remoteDispatchDomainSetMemoryParameters (server=0x555555eb7e00,
client=0x555555ec4b10, msg=0x555555eb94e0, rerr=0x7fffec8bfb70, args=0x7fffe40b8510)
at remote_dispatch.h:7621
#8 0x00005555556133fd in remoteDispatchDomainSetMemoryParametersHelper (server=0x555555eb7e00,
client=0x555555ec4b10, msg=0x555555eb94e0, rerr=0x7fffec8bfb70, args=0x7fffe40b8510,
ret=0x7fffe40b84f0) at remote_dispatch.h:7591
#9 0x00007ffff73b293f in virNetServerProgramDispatchCall (prog=0x555555ec3ae0,
server=0x555555eb7e00, client=0x555555ec4b10, msg=0x555555eb94e0)
at rpc/virnetserverprogram.c:435
#10 0x00007ffff73b207f in virNetServerProgramDispatch (prog=0x555555ec3ae0,
server=0x555555eb7e00, client=0x555555ec4b10, msg=0x555555eb94e0)
at rpc/virnetserverprogram.c:305
#11 0x00007ffff73a4d2c in virNetServerProcessMsg (srv=0x555555eb7e00, client=0x555555ec4b10,
prog=0x555555ec3ae0, msg=0x555555eb94e0) at rpc/virnetserver.c:165
#12 0x00007ffff73a4e8d in virNetServerHandleJob (jobOpaque=0x555555ec3e30, opaque=0x555555eb7e00)
at rpc/virnetserver.c:186
#13 0x00007ffff7187f3f in virThreadPoolWorker (opaque=0x555555eb7ac0) at util/virthreadpool.c:144
#14 0x00007ffff718733a in virThreadHelper (data=0x555555eb7890) at util/virthreadpthread.c:161
#15 0x00007ffff468ed89 in start_thread (arg=0x7fffec8c0700) at pthread_create.c:308
#16 0x00007ffff3da26bd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:113
CVE-2013-6436: fix crash in lxcDomainGetMemoryParameters
The function doesn't check whether the request is made for active or
inactive domain. Thus when the domain is not running it still tries
accessing non-existing cgroups (priv->cgroup, which is NULL).
I re-made the function in order for it to work the same way it's qemu
counterpart does.
Reproducer:
1) Define an LXC domain
2) Do 'virsh memtune <domain>'
Backtrace:
Thread 6 (Thread 0x7fffec8c0700 (LWP 13387)):
#0 0x00007ffff70edcc4 in virCgroupPathOfController (group=0x0, controller=3,
key=0x7ffff75734bd "memory.limit_in_bytes", path=0x7fffec8bf750) at util/vircgroup.c:1764
#1 0x00007ffff70e958c in virCgroupGetValueStr (group=0x0, controller=3,
key=0x7ffff75734bd "memory.limit_in_bytes", value=0x7fffec8bf7c0) at util/vircgroup.c:705
#2 0x00007ffff70e9d29 in virCgroupGetValueU64 (group=0x0, controller=3,
key=0x7ffff75734bd "memory.limit_in_bytes", value=0x7fffec8bf810) at util/vircgroup.c:804
#3 0x00007ffff70ee706 in virCgroupGetMemoryHardLimit (group=0x0, kb=0x7fffec8bf8a8)
at util/vircgroup.c:1962
#4 0x00005555557d590f in lxcDomainGetMemoryParameters (dom=0x7fffd40024a0,
params=0x7fffd40027a0, nparams=0x7fffec8bfa24, flags=0) at lxc/lxc_driver.c:826
#5 0x00007ffff72c28d3 in virDomainGetMemoryParameters (domain=0x7fffd40024a0,
params=0x7fffd40027a0, nparams=0x7fffec8bfa24, flags=0) at libvirt.c:4137
#6 0x000055555563714d in remoteDispatchDomainGetMemoryParameters (server=0x555555eb7e00,
client=0x555555ebaef0, msg=0x555555ebb3e0, rerr=0x7fffec8bfb70, args=0x7fffd40024e0,
ret=0x7fffd4002420) at remote.c:1895
#7 0x00005555556052c4 in remoteDispatchDomainGetMemoryParametersHelper (server=0x555555eb7e00,
client=0x555555ebaef0, msg=0x555555ebb3e0, rerr=0x7fffec8bfb70, args=0x7fffd40024e0,
ret=0x7fffd4002420) at remote_dispatch.h:4050
#8 0x00007ffff73b293f in virNetServerProgramDispatchCall (prog=0x555555ec3ae0,
server=0x555555eb7e00, client=0x555555ebaef0, msg=0x555555ebb3e0)
at rpc/virnetserverprogram.c:435
#9 0x00007ffff73b207f in virNetServerProgramDispatch (prog=0x555555ec3ae0,
server=0x555555eb7e00, client=0x555555ebaef0, msg=0x555555ebb3e0)
at rpc/virnetserverprogram.c:305
#10 0x00007ffff73a4d2c in virNetServerProcessMsg (srv=0x555555eb7e00, client=0x555555ebaef0,
prog=0x555555ec3ae0, msg=0x555555ebb3e0) at rpc/virnetserver.c:165
#11 0x00007ffff73a4e8d in virNetServerHandleJob (jobOpaque=0x555555ebc7e0, opaque=0x555555eb7e00)
at rpc/virnetserver.c:186
#12 0x00007ffff7187f3f in virThreadPoolWorker (opaque=0x555555eb7ac0) at util/virthreadpool.c:144
#13 0x00007ffff718733a in virThreadHelper (data=0x555555eb7890) at util/virthreadpthread.c:161
#14 0x00007ffff468ed89 in start_thread (arg=0x7fffec8c0700) at pthread_create.c:308
#15 0x00007ffff3da26bd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:113
Eric Blake [Tue, 10 Dec 2013 12:45:26 +0000 (05:45 -0700)]
storage: fix omitted slash in gluster volume URI
When doing 'virsh vol-dumpxml' on a gluster pool's volume, the
resulting URI incorrectly omitted a slash between hostname and
path: gluster://192.168.122.206rhsvol1/fedora-19.img
This is fallout from me rebasing earlier versions of my patch
that ended up as commit efee1af; I had originally played with
always requiring the gluster volume to have a leading slash,
but it was easier to use the gluster API if the gluster volume
name was guaranteed to have no slash. While I got the URI of
the pool correct, I forgot to fix the URI of a libvirt volume.
* src/storage/storage_backend_gluster.c
(virStorageBackendGlusterRefreshVol): Use correct starting point
since uri construction requires leading slash.
Ryota Ozaki [Sun, 1 Dec 2013 14:46:06 +0000 (23:46 +0900)]
vbox: handle errors of virDomainHostdevDefAlloc correctly
The original code ignored errors of virDomainHostdevDefAlloc,
however, we should properly do error return from the function
if it occurs.
The fix pulls out virDomainHostdevDefAlloc from the loop and
executes it all together before the loop. So we can easily
return on errors without the notion of other memory allocations
in the loop.
The deallocation code is separated from the allocation code
because it will be used by a further patch for fixing other error
handlings.
Reported-by: Laine Stump <laine@laine.org> Signed-off-by: Ryota Ozaki <ozaki.ryota@gmail.com>
Ryota Ozaki [Sun, 1 Dec 2013 14:46:05 +0000 (23:46 +0900)]
vbox: fix incorrect loop condition in vboxHostDeviceGetXMLDesc
The fixed loop used logical OR to combine two conditions, however,
it is apparently incorrect and logical AND is correct.
We can fix it by replacing OR with AND, but this patch instead
fixes the problem by getting rid of the first conditional
statement: USBFilterCount < def->nhostdevs. It isn't needed
because USBFilterCount will never be greater than or equal to
def->nhostdevs.
def->nhostdevs is calculated in the following code
above the loop in question like this:
for (i = 0; i < deviceFilters.count; i++) {
PRBool active = PR_FALSE;
IUSBDeviceFilter *deviceFilter = deviceFilters.items[i];
deviceFilter->vtbl->GetActive(deviceFilter, &active);
if (active) {
def->nhostdevs++;
}
}
And the loop is constructed as like this:
for (i = 0; (USBFilterCount < def->nhostdevs) || (i < deviceFilters.count); i++) {
PRBool active = PR_FALSE;
(snip)
deviceFilter->vtbl->GetActive(deviceFilter, &active);
if (!active)
continue;
(snip)
USBFilterCount++;
}
So def->nhostdevs is the number of active device filters and
USBFilterCount is counted up only when a device filter is active.
Thus, we can remove USBFilterCount < def->nhostdevs safely.
Reported-by: Laine Stump <laine@laine.org> Signed-off-by: Ryota Ozaki <ozaki.ryota@gmail.com>
While running nwfilterxml2xmltest, it was found that valgrind pointed out the
following error...
==7466== 16 bytes in 1 blocks are definitely lost in loss record 26 of 90
==7466== at 0x4A06B6F: calloc (vg_replace_malloc.c:593)
==7466== by 0x4C651AD: virAlloc (viralloc.c:142)
==7466== by 0x4D0450D: virNWFilterDefParseNode (nwfilter_conf.c:2575)
==7466== by 0x4D05D84: virNWFilterDefParse (nwfilter_conf.c:2647)
==7466== by 0x401FDE: testCompareXMLToXMLHelper (nwfilterxml2xmltest.c:39)
==7466== by 0x402DE1: virtTestRun (testutils.c:138)
==7466== by 0x4018E9: mymain (nwfilterxml2xmltest.c:111)
==7466== by 0x403482: virtTestMain (testutils.c:593)
==7466== by 0x341F421A04: (below main) (libc-start.c:225)
...21 times, which are related to 21 tests in nwfilterxml2xmltest.c which sent
EXPECT_WARN = false. There were two scenarios in virNWFilterDefParseXML(),
when the variable 'entry' was malloc'ed, but not freed.
This patch fixes the memory leaks found while running qemuxml2argvtest
==8260== 3 bytes in 1 blocks are definitely lost in loss record 1 of
129
==8260== at 0x4A0887C: malloc (vg_replace_malloc.c:270)
==8260== by 0x341F485E21: strdup (strdup.c:42)
==8260== by 0x4CADCFF: virStrdup (virstring.c:554)
==8260== by 0x4CBB839: virXPathString (virxml.c:90)
==8260== by 0x4CE753A: virDomainDefParseXML (domain_conf.c:11478)
==8260== by 0x4CEB4FE: virDomainDefParseNode (domain_conf.c:12742)
==8260== by 0x4CEB675: virDomainDefParse (domain_conf.c:12684)
==8260== by 0x425958: testCompareXMLToArgvHelper (qemuxml2argvtest.c:107)
==8260== by 0x427111: virtTestRun (testutils.c:138)
==8260== by 0x41D3FE: mymain (qemuxml2argvtest.c:452)
==8260== by 0x4277B2: virtTestMain (testutils.c:593)
==8260== by 0x341F421A04: (below main) (libc-start.c:225)
==8260==
==8260== 4 bytes in 1 blocks are definitely lost in loss record 5 of
129
==8260== at 0x4A0887C: malloc (vg_replace_malloc.c:270)
==8260== by 0x341F485E21: strdup (strdup.c:42)
==8260== by 0x4CADCFF: virStrdup (virstring.c:554)
==8260== by 0x4CBB839: virXPathString (virxml.c:90)
==8260== by 0x4CE753A: virDomainDefParseXML (domain_conf.c:11478)
==8260== by 0x4CEB4FE: virDomainDefParseNode (domain_conf.c:12742)
==8260== by 0x4CEB675: virDomainDefParse (domain_conf.c:12684)
==8260== by 0x425958: testCompareXMLToArgvHelper (qemuxml2argvtest.c:107)
==8260== by 0x427111: virtTestRun (testutils.c:138)
==8260== by 0x41D39A: mymain (qemuxml2argvtest.c:451)
==8260== by 0x4277B2: virtTestMain (testutils.c:593)
==8260== by 0x341F421A04: (below main) (libc-start.c:225)
==8260==
The basic problem is that during a network update, the required
iptables rules sometimes change, and this was being handled by simply
removing and re-adding the rules. However, the removal of the old
rules was done based on the *new* state of the network, which would
mean that some of the rules would not match those currently in the
system, so the old rules wouldn't be removed.
This patch removes the old rules prior to updating the network
definition then adds the new rules as soon as the definition is
updated. Note that this could lead to a stray packet or two during the
interim, but that was already a problem before (the period of limbo is
now just slightly longer).
While moving the location for the rules, I added a few more sections
that should result in the iptables rules being redone:
DHCP_RANGE and DHCP_HOST - these are needed because adding/removing a dhcp
host entry could lead to the dhcp service being started/stopped, which
would require that the mangle rule that fixes up dhcp response
checksums sould need to be added/removed, and this wasn't being done.
Eric Blake [Wed, 27 Nov 2013 21:59:52 +0000 (14:59 -0700)]
tests: fix virpcitest with read-only srcdir
'make distcheck' has been broken since commit 21685c9; basically,
it emulates the case of a read-only $(srcdir) (such as building
from a tarball exploded onto a CD-ROM), but we were creating our
fake pci device as a symlink into $(srcdir) and failing when that
requires opening the config file for writing:
3) testVirPCIDeviceReset ... libvirt: error : Failed to open config space file '/sys/bus/pci/devices/0000:00:01.0/config': Permission denied
Fix it by copying rather than symlinking.
* tests/virpcimock.c (make_file): Add parameter to allow binary
creation; adjust all callers.
(pci_device_new_from_stub): Copy rather than symlink.
Eric Blake [Wed, 27 Nov 2013 21:31:53 +0000 (14:31 -0700)]
tests: guarantee abs_srcdir in all C tests
While trying to debug a failure of virpcitest during 'make distcheck',
I noticed that with a VPATH build, 'cd tests; ./virpcitest' fails for
an entirely different reason. To reproduce the distcheck failure, I
had to run 'cd tests; abs_srcdir=/path/to/src ./virpcitest'. But we
document in HACKING that all of our tests are supposed to be runnable
without requiring extra environment variables.
The solution: hardcode the location of srcdir into the just-built
binaries, rather than requiring make to prepopulate environment
variables. With this, './virpcitest' passes even in a VPATH build
(provided that $(srcdir) is writable; a followup patch will fix the
conditions required by 'make distcheck'). [Note: the makefile must
still pass on directory variables to the test environment of shell
scripts, since those aren't compiled. So while this solves the case
of a compiled test, it still requires environment variables to pass
a VPATH build of any shell script test case that relies on srcdir.]
* tests/Makefile.am (AM_CFLAGS): Define abs_srcdir in all compiled
tests.
* tests/testutils.h (abs_srcdir): Quit declaring.
* tests/testutils.c (virtTestMain): Rely on define rather than
environment variable.
* tests/virpcimock.c (pci_device_new_from_stub): Rely on define.
* tests/cputest.c (mymain): Adjust abs_top_srcdir default.
* tests/qemuxml2argvtest.c (mymain): Likewise.
* tests/qemuxmlnstest.c (mymain): Likewise.
Eric Blake [Wed, 27 Nov 2013 03:57:05 +0000 (20:57 -0700)]
storage: skip selinux cleanup when fd not available
When attempting to backport gluster pools to an older versoin
where there is no VIR_STRDUP, I got a crash from calling
strdup(,NULL). Rather than relying on the current else branch
safely doing nothing when there is no fd, it is easier to just
skip it. While at it, there's no need to explicitly set
perms.label to NULL after a VIR_FREE().
* src/storage/storage_backend.c
(virStorageBackendUpdateVolTargetInfoFD): Minor optimization.
Bing Bu Cao [Wed, 27 Nov 2013 10:52:12 +0000 (18:52 +0800)]
qemu: preserve netdev MAC address during 'domxml-to-native'
The virsh command 'domxml-to-native' (virConnectDomainXMLToNative())
converts all network devices to "type='ethernet'" in order to make it
more likely that the generated command could be run directly from a
shell (other libvirt network device types end up referencing file
descriptors for tap devices assumed to have been created by libvirt,
which can't be done in this case).
During this conversion, all of the netdev parameters are cleared out,
then specific items are filled in after changing the type. The MAC
address was not one of these preserved items, and the result was that
mac addresses in the generated commandlines were always
00:00:00:00:00:00.
This patch saves the mac address before the conversion, then
repopulates it afterwards, so the proper mac addresses show up in the
commandline.
Signed-off-by: Bing Bu Cao <mars@linux.vnet.ibm.com> Signed-off-by: Laine Stump <laine@laine.org>
Eric Blake [Mon, 25 Nov 2013 21:38:22 +0000 (14:38 -0700)]
storage: don't read storage volumes in nonblock mode
Commit 348b4e2 introduced a potential problem (thankfully not
in any release): we are attempting to use virFileReadHeaderFD()
on a file that was opened with O_NONBLOCK. While this
shouldn't be a problem in practice (because O_NONBLOCK
typically doesn't affect regular or block files, and fifos and
sockets cannot be storage volumes), it's better to play it safe
to avoid races from opening an unexpected file type while also
avoiding problems with having to handle EAGAIN while read()ing.
Based on a report by Dan Berrange.
* src/storage/storage_backend.c
(virStorageBackendVolOpenCheckMode): Fix up fd after avoiding race.
unprivileged user has no rights to move the mounts that
inherited from parent mountns. we use this feature to move
the /stateDir/domain-name.{dev, devpts} to the /dev/ and
/dev/pts directroy of container. this commit breaks libvirt lxc.
this patch changes the behavior to bind these mounts when
user namespace is enabled and move these mounts when user
namespace is disabled.
sasl: Fix authentication when using PLAIN mechanism
With some authentication mechanism (PLAIN for example), sasl_client_start()
can return SASL_OK, which translates to virNetSASLSessionClientStart()
returning VIR_NET_SASL_COMPLETE.
cyrus-sasl documentation is a bit vague as to what to do in such situation,
but upstream clarified this a bit in
http://asg.andrew.cmu.edu/archive/message.php?mailbox=archive.cyrus-sasl&msg=10104
When we got VIR_NET_SASL_COMPLETE after virNetSASLSessionClientStart() and
if the remote also tells us that authentication is complete, then we should
end the authentication procedure rather than forcing a call to
virNetSASLSessionClientStep(). Without this patch, when trying to use SASL
PLAIN, I get:
error :authentication failed : Failed to step SASL negotiation: -1
(SASL(-1): generic failure: Unable to find a callback: 32775)
This patch is based on a spice-gtk patch by Dietmar Maurer.
Fix invalid read in virNetSASLSessionClientStep debug log
virNetSASLSessionClientStep logs the data that is going to be passed to
sasl_client_step as input data. However, it tries to log it as a string,
while there is no guarantee that this data is going to be nul-terminated.
This leads to this valgrind log:
==20938== Invalid read of size 1
==20938== at 0x8BDB08F: vfprintf (vfprintf.c:1635)
==20938== by 0x8C06DF2: vasprintf (vasprintf.c:62)
==20938== by 0x4CCEDF9: virVasprintfInternal (virstring.c:337)
==20938== by 0x4CA9516: virLogVMessage (virlog.c:842)
==20938== by 0x4CA939A: virLogMessage (virlog.c:778)
==20938== by 0x4E21E0D: virNetSASLSessionClientStep (virnetsaslcontext.c:458)
==20938== by 0x4DE47B8: remoteAuthSASL (remote_driver.c:4136)
==20938== by 0x4DE33AE: remoteAuthenticate (remote_driver.c:3635)
==20938== by 0x4DDBFAA: doRemoteOpen (remote_driver.c:832)
==20938== by 0x4DDC8BA: remoteConnectOpen (remote_driver.c:1027)
==20938== by 0x4D8595F: do_open (libvirt.c:1239)
==20938== by 0x4D863F3: virConnectOpenAuth (libvirt.c:1481)
==20938== by 0x12762B: vshReconnect (virsh.c:337)
==20938== by 0x12C9B0: vshInit (virsh.c:2470)
==20938== by 0x12E9A5: main (virsh.c:3338)
==20938== Address 0xe329ccd is 0 bytes after a block of size 141 alloc'd
==20938== at 0x4A081D4: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==20938== by 0x8CB91B4: xdr_array (xdr_array.c:94)
==20938== by 0x4E039C2: xdr_remote_auth_sasl_start_ret (remote_protocol.c:3134)
==20938== by 0x4E1F8AA: virNetMessageDecodePayload (virnetmessage.c:405)
==20938== by 0x4E119F5: virNetClientProgramCall (virnetclientprogram.c:377)
==20938== by 0x4DF8141: callFull (remote_driver.c:5794)
==20938== by 0x4DF821A: call (remote_driver.c:5816)
==20938== by 0x4DE46CF: remoteAuthSASL (remote_driver.c:4112)
==20938== by 0x4DE33AE: remoteAuthenticate (remote_driver.c:3635)
==20938== by 0x4DDBFAA: doRemoteOpen (remote_driver.c:832)
==20938== by 0x4DDC8BA: remoteConnectOpen (remote_driver.c:1027)
==20938== by 0x4D8595F: do_open (libvirt.c:1239)
==20938== by 0x4D863F3: virConnectOpenAuth (libvirt.c:1481)
==20938== by 0x12762B: vshReconnect (virsh.c:337)
==20938== by 0x12C9B0: vshInit (virsh.c:2470)
==20938== by 0x12E9A5: main (virsh.c:3338)
Tie SASL callbacks lifecycle to virNetSessionSASLContext
The array of sasl_callback_t callbacks which is passed to sasl_client_new()
must be kept alive as long as the created sasl_conn_t object is alive as
cyrus-sasl uses this structure internally for things like logging, so
the memory used for callbacks must only be freed after sasl_dispose() has
been called.
During testing of successful SASL logins with
virsh -c qemu+tls:///system list --all
I've been getting invalid read reports from valgrind
==9237== Invalid read of size 8
==9237== at 0x6E93B6F: _sasl_getcallback (common.c:1745)
==9237== by 0x6E95430: _sasl_log (common.c:1850)
==9237== by 0x16593D87: digestmd5_client_mech_dispose (digestmd5.c:4580)
==9237== by 0x6E91653: client_dispose (client.c:332)
==9237== by 0x6E9476A: sasl_dispose (common.c:851)
==9237== by 0x4E225A1: virNetSASLSessionDispose (virnetsaslcontext.c:678)
==9237== by 0x4CBC551: virObjectUnref (virobject.c:262)
==9237== by 0x4E254D1: virNetSocketDispose (virnetsocket.c:1042)
==9237== by 0x4CBC551: virObjectUnref (virobject.c:262)
==9237== by 0x4E2701C: virNetSocketEventFree (virnetsocket.c:1794)
==9237== by 0x4C965D3: virEventPollCleanupHandles (vireventpoll.c:583)
==9237== by 0x4C96987: virEventPollRunOnce (vireventpoll.c:652)
==9237== by 0x4C94730: virEventRunDefaultImpl (virevent.c:274)
==9237== by 0x12C7BA: vshEventLoop (virsh.c:2407)
==9237== by 0x4CD3D04: virThreadHelper (virthreadpthread.c:161)
==9237== by 0x7DAEF32: start_thread (pthread_create.c:309)
==9237== by 0x8C86EAC: clone (clone.S:111)
==9237== Address 0xe2d61b0 is 0 bytes inside a block of size 168 free'd
==9237== at 0x4A07577: free (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==9237== by 0x4C73827: virFree (viralloc.c:580)
==9237== by 0x4DE4BC7: remoteAuthSASL (remote_driver.c:4219)
==9237== by 0x4DE33D0: remoteAuthenticate (remote_driver.c:3639)
==9237== by 0x4DDBFAA: doRemoteOpen (remote_driver.c:832)
==9237== by 0x4DDC8DC: remoteConnectOpen (remote_driver.c:1031)
==9237== by 0x4D8595F: do_open (libvirt.c:1239)
==9237== by 0x4D863F3: virConnectOpenAuth (libvirt.c:1481)
==9237== by 0x12762B: vshReconnect (virsh.c:337)
==9237== by 0x12C9B0: vshInit (virsh.c:2470)
==9237== by 0x12E9A5: main (virsh.c:3338)
This commit changes virNetSASLSessionNewClient() to take ownership of the SASL
callbacks. Then we can free them in virNetSASLSessionDispose() after the corresponding
sasl_conn_t has been freed.
remote: Don't leak priv->tls object on connection failure
When testing SASL authentication over TLS with
virsh -c qemu+tls:///system list --all
I got this valgrind trace after entering wrong credentials:
==30540== 26,903 (88 direct, 26,815 indirect) bytes in 1 blocks are definitely lost in loss record 289 of 293
==30540== at 0x4A081D4: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==30540== by 0x4C7379A: virAllocVar (viralloc.c:558)
==30540== by 0x4CBC178: virObjectNew (virobject.c:190)
==30540== by 0x4CBC329: virObjectLockableNew (virobject.c:216)
==30540== by 0x4E2D003: virNetTLSContextNew (virnettlscontext.c:719)
==30540== by 0x4E2DC3F: virNetTLSContextNewPath (virnettlscontext.c:930)
==30540== by 0x4E2DD5B: virNetTLSContextNewClientPath (virnettlscontext.c:957)
==30540== by 0x4DDB618: doRemoteOpen (remote_driver.c:627)
==30540== by 0x4DDC8BA: remoteConnectOpen (remote_driver.c:1031)
==30540== by 0x4D8595F: do_open (libvirt.c:1239)
==30540== by 0x4D863F3: virConnectOpenAuth (libvirt.c:1481)
==30540== by 0x12762B: vshReconnect (virsh.c:337)
==30540== by 0x12C9B0: vshInit (virsh.c:2470)
==30540== by 0x12E9A5: main (virsh.c:3338)
Eric Blake [Tue, 12 Nov 2013 04:08:27 +0000 (21:08 -0700)]
storage: probe qcow2 volumes in gluster pool
Putting together pieces from previous patches, it is now possible
for 'virsh vol-dumpxml --pool gluster volname' to report metadata
about a qcow2 file stored on gluster. The backing file is still
treated as raw; to fix that, more patches are needed to make the
storage backing chain analysis recursive rather than halting at
a network protocol name, but that work will not need any further
calls into libgfapi so much as just reusing this code, and that
should be the only code outside of the storage driver that needs
any help from libgfapi. Any additional use of libgfapi within
libvirt should only be needed for implementing storage pool APIs
such as volume creation or resizing, where backing chain analysis
should be unaffected.
* src/storage/storage_backend_gluster.c
(virStorageBackendGlusterReadHeader): New helper function.
(virStorageBackendGlusterRefreshVol): Probe non-raw files.
Eric Blake [Mon, 18 Nov 2013 22:24:05 +0000 (15:24 -0700)]
storage: improve handling of symlinks in gluster
With this patch, dangling and looping symlinks are silently
ignored, while links to files and directories are treated the
same as the underlying file or directory. This is the same
behavior as both 'directory' and 'netfs' pools.
* src/storage/storage_backend_gluster.c
(virStorageBackendGlusterRefreshVol): Treat symlinks similar to
directory and netfs pools.
Eric Blake [Mon, 18 Nov 2013 19:45:31 +0000 (12:45 -0700)]
storage: improve allocation stats reported on gluster files
We already had code for handling allocation different than
capacity for sparse files; we just had to wire it up to be
used when inspecting gluster images.
Eric Blake [Wed, 20 Nov 2013 20:17:55 +0000 (13:17 -0700)]
storage: improve directory support in gluster pool
Take advantage of the previous patch's addition of 'netdir' as
a distinct volume type, to expose rather than silently skip
directories embedded in a gluster pool. Also serves as an XML
validation for the previous patch.
Eric Blake [Mon, 18 Nov 2013 23:43:06 +0000 (16:43 -0700)]
storage: add network-dir as new storage volume type
In the 'directory' and 'netfs' storage pools, a user can see
both 'file' and 'dir' storage volume types, to know when they
can descend into a subdirectory. But in a network-based storage
pool, such as the upcoming 'gluster' pool, we use 'network'
instead of 'file', and did not have any counterpart for a
directory until this patch. Adding a new volume type
'network-dir' is better than reusing 'dir', because it makes
it clear that the only way to access 'network' volumes within
that container is through the network mounting (leaving 'dir'
for something accessible in the local file system).
* include/libvirt/libvirt.h.in (virStorageVolType): Expand enum.
* docs/formatstorage.html.in: Document it.
* docs/schemasa/storagevol.rng (vol): Allow new value.
* src/conf/storage_conf.c (virStorageVol): Use new value.
* src/qemu/qemu_command.c (qemuBuildVolumeString): Fix client.
* src/qemu/qemu_conf.c (qemuTranslateDiskSourcePool): Likewise.
* tools/virsh-volume.c (vshVolumeTypeToString): Likewise.
* src/storage/storage_backend_fs.c
(virStorageBackendFileSystemVolDelete): Likewise.
Eric Blake [Wed, 30 Oct 2013 03:28:16 +0000 (21:28 -0600)]
storage: implement rudimentary glusterfs pool refresh
Actually put gfapi to use, by allowing the creation of a gluster
pool. Right now, all volumes are treated as raw and directories
are skipped; further patches will allow peering into files to
allow for qcow2 files and backing chains, and reporting proper
volume allocation. This implementation was tested against Fedora
19's glusterfs 3.4.1; it might be made simpler by requiring a
higher minimum, and/or require more hacks to work with a lower
minimum.
Eric Blake [Tue, 15 Oct 2013 23:06:18 +0000 (17:06 -0600)]
storage: document gluster pool
Add support for a new <pool type='gluster'>, similar to
RBD and Sheepdog. Terminology wise, a gluster volume
forms a libvirt storage pool, within the gluster volume,
individual files are treated as libvirt storage volumes.
* docs/schemas/storagepool.rng (poolgluster): New pool type.
* docs/formatstorage.html.in: Document gluster.
* docs/storage.html.in: Likewise, and contrast it with netfs.
* tests/storagepoolxml2xmlin/pool-gluster.xml: New test.
* tests/storagepoolxml2xmlout/pool-gluster.xml: Likewise.
* tests/storagepoolxml2xmltest.c (mymain): Likewise.
Eric Blake [Tue, 19 Nov 2013 23:26:05 +0000 (16:26 -0700)]
storage: initial support for linking with libgfapi
We support gluster volumes in domain XML, so we also ought to
support them as a storage pool. Besides, a future patch will
want to take advantage of libgfapi to handle the case of a
gluster device holding qcow2 rather than raw storage, and for
that to work, we need a storage backend that can read gluster
storage volume contents. This sets up the framework.
Note that the new pool is named 'gluster' to match a
<disk type='network'><source protocol='gluster'> image source
already supported in a <domain>; it does NOT match the
<pool type='netfs'><source><target type='glusterfs'>,
since that uses a FUSE mount to a local file name rather than
a network name.
This and subsequent patches have been tested against glusterfs
3.4.1 (available on Fedora 19); there are likely bugs in older
versions that may prevent decent use of gfapi, so this patch
enforces the minimum version tested. A future patch may lower
the minimum. On the other hand, I hit at least two bugs in
3.4.1 that will be fixed in 3.5/3.4.2, where it might be worth
raising the minimum: glfs_readdir is nicer to use than
glfs_readdir_r [1], and glfs_fini should only return failure on
an actual failure [2].
* configure.ac (WITH_STORAGE_GLUSTER): New conditional.
* m4/virt-gluster.m4: new file.
* libvirt.spec.in (BuildRequires): Support gluster in spec file.
* src/conf/storage_conf.h (VIR_STORAGE_POOL_GLUSTER): New pool
type.
* src/conf/storage_conf.c (poolTypeInfo): Treat similar to
sheepdog and rbd.
(virStoragePoolDefFormat): Don't output target for gluster.
* src/storage/storage_backend_gluster.h: New file.
* src/storage/storage_backend_gluster.c: Likewise.
* po/POTFILES.in: Add new file.
* src/storage/storage_backend.c (backends): Register new type.
* src/Makefile.am (STORAGE_DRIVER_GLUSTER_SOURCES): Build new files.
* src/storage/storage_backend.h (_virStorageBackend): Documet
assumption.
Eric Blake [Tue, 19 Nov 2013 20:14:54 +0000 (13:14 -0700)]
storage: expose volume meta-type in XML
I got annoyed at having to use both 'virsh vol-list $pool --details'
AND 'virsh vol-dumpxml $vol $pool' to learn if I had populated
the volume correctly. Since two-thirds of the data present in
virStorageVolGetInfo() already appears in virStorageVolGetXMLDesc(),
this just adds the remaining piece of information, as:
<volume type='...'>
...
</volume>
* docs/formatstorage.html.in: Document new <volume type=...>.
* docs/schemas/storagevol.rng (vol): Add it to RelaxNG.
* src/conf/storage_conf.h (virStorageVolTypeToString): Declare.
* src/conf/storage_conf.c (virStorageVolTargetDefFormat): Output
the metatype.
(virStorageVolDefParseXML): Parse it, for unit tests.
* tests/storagevolxml2xmlout/vol-*.xml: Update tests to match.
Jiri Denemark [Mon, 25 Nov 2013 15:37:32 +0000 (16:37 +0100)]
spec: Don't save/restore running VMs on libvirt-client update
The previous attempt (commit d65e0e1) removed just one of two
libvirt-guests restarts that happened on libvirt-client update. Let's
remove the last one too :-)
virsh domxml-from-native to treat SCSI as the bus type for pseries by default
The bus type IDE being enum Zero, the bus type on pseries system appears as IDE for all the -hda/-cdrom and for disk drives with if="none" type. Pseries platform needs this to appear as SCSI instead of IDE. The ide being not supported, the explicit requests for ide devices will return an error.
Signed-off-by: Shivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
Eric Blake [Fri, 22 Nov 2013 19:40:15 +0000 (12:40 -0700)]
storage: allow interleave in volume XML
The RNG grammar did not allow arbitrary interleaving, which makes
it harder than necessary to create a new volume from handwritten XML.
(Compare also to commit caf516db for pools).
* docs/schemas/storagevol.rng: Support interleaving.
* tests/storagevolxml2xmlin/vol-file-backing.xml: Test it.
Ryota Ozaki [Thu, 21 Nov 2013 14:41:07 +0000 (23:41 +0900)]
vbox: add support for 4.3 APIs
Makefile.am, vbox_V4_3.c and vbox_driver.c do regular
modifitions to support a new version of APIs.
vbox_tmpl.c basically fixes incompatibilities since 4.2.
The affected incompatibilities of 4.3 are:
* IMachine::Delete() has been renamed to IMachine::deleteConfig()
* IMedium::CreateBaseStorage() now accepts multiple variant values
* IDisplay::GetScreenResolution() now returns the display position
in the guest
* IMachine now has multiple IUSBControllers and IUSBDeviceFilters
handles USB device filters instead of (obsolete) IUSBController
This patch is tested on Mac OS X 10.8.5 and Fedora 19.
Ryota Ozaki [Thu, 21 Nov 2013 14:41:06 +0000 (23:41 +0900)]
vbox: import vbox_CAPI_v4_3.h from SDK
vbox_CAPI_v4_3.h is almost same as
sdk/bindings/xpcom/include/VBoxCAPI_v4_3.h of
http://download.virtualbox.org/virtualbox/4.3.2/VirtualBoxSDK-4.3.2-90405.zip,
but modified to fix preprocessor indentations by using cppi.
Ryota Ozaki [Thu, 21 Nov 2013 14:41:05 +0000 (23:41 +0900)]
vbox: pull vboxHostDeviceGetXMLDesc out from vboxDomainGetXMLDesc
The USB-related code in vboxDomainGetXMLDesc is deeply nested and
difficult to add new code. So flatten it. To do so, the code is
pulled out from vboxDomainGetXMLDesc to make the function short
and to leaverage early return and goto for error handling.
The original problem was that libvirt_virConnectGetCPUModelNames
was listed twice in the exports table, once automatically from
the generator and once from the manual override. We merely needed
to list it in the skip_impl list, and not delete the manually
written code entirely.
Ján Tomko [Tue, 19 Nov 2013 17:03:21 +0000 (18:03 +0100)]
Don't start a nested job in qemuMigrationPrepareAny
This nested job is canceled by the first ExitMonitor call (even though
it was not created by the corresponding EnterMonitor call), and
again in qemuMigrationPrepareAny if qemuProcessStart failed.
This can lead to a crash if the vm object was disposed of before calling
qemuDomainRemoveInactive:
0 ..62bc in virClassIsDerivedFrom (klass=0xdeadbeef,
parent=0x7ffce4cdd270) at util/virobject.c:166
1 ..6666 in virObjectIsClass at util/virobject.c:362
2 ..66b4 in virObjectLock at util/virobject.c:314
3 ..477e in virDomainObjListRemove at conf/domain_conf.c:2359
4 ..7a64 in qemuDomainRemoveInactive at qemu/qemu_domain.c:2087
5 ..956c in qemuMigrationPrepareAny at qemu/qemu_migration.c:2469
Which in a default configuration will managedsave every running VM,
and then restore them. Certainly not something we should do every
time the libvirt-client RPM is updated.
Just drop the try-restart attempt, I don't know what purpose it
serves anyways.
As virt-login-shell is an SUID binary, we should restrict its usage to
just the users chosen by an administrator to use virt-login-shell as
their login shell. This can easily be done by making the binary
executable only by users from a new virtlogin group.
If a domain is rebooting and a migrate API is called meanwhile we would
have to transfer the fakeReboot attribute to the destination in order to
prevent domain doing plain shutdown over there. We shouldn't try to do
anything clever about it other than documenting this as a known
limitation.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
build: Don't fail on '<' or '>' with old xmllint
Older xmllint version don't allow such characters in datatype anyURI.
In order not to change too much, I'm suggesting making a choice of
anyURI or 'absPathName' which should be fine (checked with upstream
and that old xmllint, both work fine).
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Oops. That's not valid XML. And when we fix the XML
generation, it fails RelaxNG validation.
I'm also tired of seeing <key>(null)</key> in the example
output for volume xml; while we used NULLSTR() to avoid
a NULL deref rather than relying on glibc's printf
extension behavior, it's even better if we avoid the issue
in the first place. But this requires being careful that
we don't invalidate any storage backends that were relying
on key being unassigned during virStoragVolCreateXML[From].
I would have split this into two patches (one for escaping,
one for avoiding <key>(null)</key>), but since they both
end up touching a lot of the same test files, I ended up
merging it into one.
Note that this patch allows pretty much any volume name
that can appear in a directory (excluding . and .. because
those are special), but does nothing to change the current
(unenforced) RelaxNG claim that pool names will consist
only of letters, numbers, _, -, and +. Tightening the C
code to match RelaxNG patterns and/or relaxing the grammar
to match the C code for pool names is a task for another
day (but remember, we DID recently tighten C code for
domain names to exclude a leading '.').
* src/conf/storage_conf.c (virStoragePoolSourceFormat)
(virStoragePoolDefFormat, virStorageVolTargetDefFormat)
(virStorageVolDefFormat): Escape user-controlled strings.
(virStorageVolDefParseXML): Parse key, for use in unit tests.
* src/storage/storage_driver.c (storageVolCreateXML)
(storageVolCreateXMLFrom): Ensure parsed key doesn't confuse
volume creation.
* docs/schemas/basictypes.rng (volName): Relax definition.
* tests/storagepoolxml2xmltest.c (mymain): Test it.
* tests/storagevolxml2xmltest.c (mymain): Likewise.
* tests/storagepoolxml2xmlin/pool-dir-naming.xml: New file.
* tests/storagepoolxml2xmlout/pool-dir-naming.xml: Likewise.
* tests/storagevolxml2xmlin/vol-file-naming.xml: Likewise.
* tests/storagevolxml2xmlout/vol-file-naming.xml: Likewise.
* tests/storagevolxml2xmlout/vol-*.xml: Fix fallout.
Doug Goldstein [Thu, 21 Nov 2013 14:47:08 +0000 (08:47 -0600)]
python: remove virConnectGetCPUModelNames from globals
Commit de51dc9c9aed0e615c8b301cccb89f4859324eb0 primarily added
virConnectGetCPUModelNames as libvirt.getCPUModelNames(conn, arch)
instead of libvirt.virConnect.getCPUModelNames(arch) so revert the code
that does the former while leaving the code that does the later.
This is the rest of the patch that was ACK'd by Dan but I committed only
the partial patch in 6a8b8ae.
Doug Goldstein [Thu, 21 Nov 2013 14:47:08 +0000 (08:47 -0600)]
python: remove virConnectGetCPUModelNames from globals
Commit de51dc9c9aed0e615c8b301cccb89f4859324eb0 primarily added
virConnectGetCPUModelNames as libvirt.getCPUModelNames(conn, arch)
instead of libvirt.virConnect.getCPUModelNames(arch) so revert the code
that does the former while leaving the code that does the later.
projects / thirdparty / libvirt.git / log
